Mortgage structure, saving rates and the wealth distribution

Luís Teles Morais

Erasmus University Rotterdam (visiting SAFE)


Goethe University Finance Brown Bag
Frankfurt am Main, 18 November 2026

Introduction

How do fixed repayment schedules in mortgages affect saving and wealth?

Mortgage debt repayments are a large fraction of household saving flows

  • Housing main asset on wealth portfolios (stock)
  • But also in flows: \(\sim 30\%\) of Euro area aggregate saving (similar in US)

Mortgages often mandate fixed repayment schedules e.g. annuity loans Amortizing vs. IO

  • Fixed at origination; deviating is costly (refinancing, late penalties, …)

New homeowners are highly constrained: hand-to-mouth, high debt and expected income growth

  • Average recent homebuyer in the Euro area: liquid wealth \(= 17.5\%\) of annual income

This paper: fixed repayments increase overall saving (\(\downarrow C\)), but crowd out liquid saving

  • Reduced consumption smoothing, increase share of hand-to-mouth
  • Effects concentrated at bottom of income and wealth distributions

This paper: model + data

Life-cycle model with costly deviation from repayment schedule

  • Homeowners face uninsurable income risk, optimize saving and repayment over mortgage lifecycle
  • HFCS data on Netherlands policy that increased cost of deviation for new homebuyers, via tax incentives

Model matches life-cycle moments of liquid wealth and mortgage debt and reveals effects of fixed repayments

  • Pre-policy, full life-cycle data identifies preferences; post-policy behavior identifies repayment friction
  • Policy equivalent to imposing a \(67\%\) premium on under-payment

Main results from applying pre-reform repayment rules to post-reform buyers in the model

  • 1. Homeowners save more, reducing consumption by \(10\%\) of income, but are more exposed to shocks
  • 2. \(\uparrow\) Total wealth/income, but \(\downarrow\) liquid wealth \(\Rightarrow\) 15 p.p. \(\uparrow\) \(\%\) HtM
    Mean liquid wealth \(\downarrow\) \(24\%\) \(\rightarrow\) \(19\%\) of annual income
    • Effects stronger at the bottom: \(\downarrow\) wealth inequality but \(\uparrow\) financial wealth inequality
  • 3. Restriction to consumption smoothing very costly: equivalent to \(\approx\) 2-3% of lifetime \(C\)

Contribution

Paper fits in literature modelling life-cycle and welfare effects of mortgage structure
Boar et al. (2022), Backman et al. (2025), Balke et al. (2024), Boutros et al. (2025), Ferreira et al. (2024), Gete & Zecchetto (2018, 2024), Greenwald et al. (2018), Karlman et al. (2021), Vestman (2018)

  • Effects of amortization requirements on household consumption and saving
    Andersen et al. (2021); Backman & Khorunzhina (2024); Backman et al. (2024); Bernstein & Koudijs (2024), Larsen et al. (2024)
    • This paper: crowding out of liquid wealth; long-run/life-cycle effects
  • Wealth distribution and the high share of “wealthy hand-to-mouth”
    Aguiar, Bils & Boar (2025), Boar, Gorea & Midrigan (2022), Kaplan & Violante (2014)
    • This paper: evidence of clear mechanism driving up WHtM share
  • Optimal mortgage contract structure
    Campbell and Cocco (2015), Campbell et al. (2018), Chambers et al. (2009), Gete (2018), Guren et al. (2018), Piskorski & Tchistyi (2010, 2011)
    • This paper: (heterogeneous) effects on household wealth and welfare of repayment rigidity

Agenda

  1. Introduction
  2. Model framework
  3. Data: stylized facts and calibration
  4. Model results
  5. Conclusion

Model framework

Model framework

Overview

Life-cycle model of consumption and saving

  • First‐time homebuyer life‐cycle, from origination to mortgage maturity
  • Two assets: liquid safe asset (risk‐free) vs. mortgage debt; housing fixed
  • Idiosyncratic income risk (permanent + transitory)
  • Realistic mortgages with repayment transaction costs, including the cost of deferring principal
    • How does this wedge affect saving and wealth accumulation?

Model framework

Household life cycle endowments and decisions

  • A home worth \(P_{0}\) (normalized) and a 30‐year fixed‐rate mortgage with balance \(M_{0}\)
  • Initial financial wealth \(A_{0}\) and exogenous risky earnings \(Y_{t}\) over the life cycle
  • Each period, choose consumption \(c_t\) and mortgage repayment \(d_t\)

Utility from non‐housing consumption only (\(H\) fixed):

\[ U(c_{t}) = \frac{c_{t}^{\,1-\gamma}}{1 - \gamma} \]

  • Separable preferences, so \(\text{argmax} \sum_t u(C_t) = \text{argmax} \sum_t u(C_t, \bar{H})\) (Campbell-Cocco 2015)

Model framework

Assets & mortgage frictions

Liquid saving and mortgage debt

  • Savings in the liquid asset (\(a_{t}\)) earn risk‐free interest \(r\)
  • Borrowing constraints:
    • \(a_{t} \ge \theta^a\,\) unsecured debt allowed and costly, paying \(r^{-}\)
    • Household cannot increase mortgage debt, only repay or maintain balance \(\,d_{t} \ge 0\,\)
  • Outstanding mortgage debt demands interest with a positive spread \(r^m > r\)

Mortgage repayment schedule

  • Mandatory amortization: \(D^{*}(m_{t-1},\,t)\) from standard annuity formula
    • Deviating from repayment schedule \(d_t < d_t^*\), then incurs transaction cost \(\tau_t ≥ 0\)
  • If default, lose house and keep low consumption \(\underline{c}\) until end

Model framework

Full dynamic household problem

The household solves, choosing consumption \(c_t\) and repayment \(d_t\):

\[ V(t, s_t) = \max_{\{c_k, d_k\}_{k=t}^{T}} \; \mathbb{E}_t \left[ \sum_{k=t}^{T-1} \beta^{k-t} \frac{c_k^{1 - \gamma}}{1 - \gamma} + \beta^{T - t} B(a_T - m_T) \right], \; \textrm{s.t.} \]

\[ \begin{aligned} a_{t+1} &= (1 + r)\bigl[a_t + y_t - (r + s)m_t - d_t - \tau_t - c_t\bigr], \\ \quad m_{t+1} &= m_t - d_t \\ \tau_t &= \tau \cdot \max\{0, d_t^* - d_t\}, \quad a_t \ge \theta^a, \quad m_t \ge 0, \quad d_t \ge 0 \end{aligned} \]

  • Terminal condition: bequest motive \(B(a_T - m_T) = b_0 \frac{(a_T - m_T + b_1)^{\,1-\gamma}}{1-\gamma}\)
  • Full mortgage repayment by retirement, with any remaining liquid wealth left as a bequest
  • Solved via deep-learning algorithm (Duarte et al. 2022; Barrera & de Silva 2024)

Data: stylized facts and calibration strategy

Data from the Netherlands and calibration strategy

HFCS: harmonized Euro-area survey, three waves (2013-21)

  • Detailed household balance sheets, income, consumption, and mortgages

A 2013 reform raised the cost of repayment flexibility for new buyers

  • Before 2013: flexible repayment, large interest-only at no extra cost
  • From 2013: interest deduction restricted to the conforming part
    • Eliminating the rebate \(\approx\) \(\uparrow\) 2x cost of carrying non-conforming debt

Initial conditions and calibration by purchase cohort

  • Households simulated by drawing from the empirical joint distribution among recent homebuyers
  • Composition stable; some drop in interest rates and house prices:
Time of purchase Pre-2013 Post-2013
LTV at origination 1.05 1.00
Liquid wealth / income -0.13 0.31
Age at purchase 32.14 32.95
House price / income 4.63 3.59
Mortgage interest rate (%) 4.56 2.36
High education (%) 69.15 73.59
  • Reduced selection concerns; model allows to account for differences at purchase

What identifies preferences and repayment costs

  • Pre-2013 buyers (free flexibility) → full life-cycle identifies preferences (\(\beta\), \(b_0\), \(\tau^{pre}\))
  • Post-2013 buyers (costly) → change in early amortization identifies the friction \(\tau^{post}\)
  • Liquid wealth of post-2013 buyers not targeted → model reproduces observed crowd-out (validation)

Cohort of post-2013 homebuyers subject to policy is repaying faster

Outstanding loan balance over time since origination

Cohort of post-2013 homebuyers is building less liquid wealth

Liquid wealth over time since origination

Calibration

Externally calibrated parameters

  • Income process: Moments of stochastic process from NL micro data (de Nardi et al. 2021) Income process
    • Two permanent income types: college vs. lower education
  • Other parameters:

Model results

Debt repayment in the model and survey

Remaining loan balance in the model and the survey

  • Post-2013, higher penalty leads to faster early repayment (even with lower int. rates)

Early post-reform liquid wealth in the model and survey

  • Early post-reform liquid wealth: 42.7% of annual income in the survey, 43.0% in the model

Liquid wealth relative to income in the model and the survey

  • Pre-2013 calibration target, model fits well
  • Post-2013 liquid wealth not targeted

Estimated parameters

  • Simulated method of moments (SMM), matching model moments to the Dutch HFCS
\(\beta\) 0.987 Annual patience
\(b_0\) 3.9 Value of liquid wealth post-retirement
\(\tau^{pre}\) 0.01 Cost of delayed repayment pre-policy (small)
\(\tau^{post}\) 0.665 Large penalty post-policy

\(\Delta\tau \approx 0.66\): effect of policy restricting repayment flexibility

  • Reflects tax penalty and potential changes in bank products
  • Large enough to overcome the low-rate boost (post-2013 rates fell below liquid returns)
  • Identified by sharp increase in early-stage repayment

Income-poorer model HHs save more in total, but less into liquid wealth

Means of model population across income quintiles (conditional on age)

Costly deferral Flexible repayment

Costly underpayment raises hand-to-mouth incidence

Model counterfactuals for the same buyers

  • Hand-to-mouth (HtM) defined as liquid wealth below one month of income
Outcome Flexible repayment Costly underpayment
Mean liquid wealth / annual income 24% 19%
Hand-to-mouth share 4% 19%

By income quintile

Hand-to-mouth rates by income quintile under the two repayment regimes

Welfare costs of mandatory amortization

Median consumption-equivalent loss

Simulated buyers Median loss
All buyers 2.41%
Lower education 2.82%
Higher education 2.13%
  • Permanent consumption increase needed to offset the welfare loss
  • Same homeowners, house prices and mortgage rates under both contracts

Forced amortization increases saving rates at the bottom of wealth dist.

Implying effects on distribution of total and financial wealth

Costly deferral Flexible repayment

  • Saving rates increase for groups at the bottom wealth groups
  • Implications for wealth distribution: \(\downarrow\) total wealth inequality but \(\uparrow\) financial w. ineq., \(\% HtM\)

Conclusion

Conclusion

  • Mortgage debt repayment is an important part of household saving flows (30\(\%\) in Euro area)
  • This paper: fixed repayment schedules crowd out liquid saving, limiting consumption smoothing
  • Imposing amortization, through a \(67\%\) penalty on delayed repayment (Netherlands), leads to:
    • Mean liquid wealth \(\downarrow\) \(24\%\) \(\rightarrow\) \(19\%\) of annual income
    • Share of hand-to-mouth \(\uparrow\) \(4\%\) \(\rightarrow\) \(19\%\)
    • Welfare loss \(\approx\) \(2–3\%\) of lifetime consumption (\(\approx\) one extra working year)
  • Distributional implications: \(\downarrow\) inequality in net worth, but \(\uparrow\) in liquid financial wealth
    • Micro-foundation for large share of “wealthy hand-to-mouth”
  • Amortization requirements may have financial stability benefits, but carry large costs for homeowners
    • Younger, lower-income households seem to be excessively penalized by rigid structure
    • Some flexibility can go a long way (optimal level for future work!)

Thank you!

Reach out: luistelesm.github.io | luis.teles.m@novasbe.pt

Appendix

Appendix

What is an interest-only mortgage?

Back to main

Appendix

Hand-to-mouth incidence by income quintile

  • Hand-to-mouth (HtM) defined as liquid wealth below one month of income
Hand-to-mouth rates by income quintile under the two repayment regimes

Back to main

Appendix

Data: mortgage amortization in the HFCS

% of regular payment going to amortization

% of household income going to amortization

Back to main

Appendix

Amortization by wave

Back to main

Appendix

Amortization for mortgages before and after 2013

Percentage of obs. where amortization is less than 5% of the regular payment:

NL others
Mortgages before 2013 30.1 1.7
Mortgages on or after 2013 11.8 1.0

Back to main

Appendix

Interest rates

Back to main

Appendix

Amortization implied by annuity formula

  • If mortgage is an annuitized loan, the amortization paid as part of the installment in period \(t\) is: \[L\times r\times\left(\frac{1}{1-\frac{1}{(1+r)^{T-t}}}-1\right)\]
    • where \(L\) is the outstanding amount, \(r\) the loan rate and \(T\) the residual maturity.
  • This is what we observe for the median HH in the overall sample but not in NL:

Back to main

Appendix

Weight of regular mortgage payments on income

Back to main

Appendix

Saving rate measure checks

  • Match with self-reported ability to save:

  • HFCS aggregates vs. national accounts (QSA)

Back to main

Appendix

Income process

Inelastic labor supply yields earnings \(Y_{t} = \Gamma_{t}Z_{t}\,\theta_{t}\), as standard (Carroll & Samwick, 1997)

  • \(\ln Z_{t} = \ln Z_{t-1} + \ln \psi_{t}\);
    \(\ln\psi_{t}\sim N\bigl(-\tfrac12\sigma_{\psi}^{2},\,\sigma_{\psi}^{2}\bigr)\) ; \(\ln \theta_{t}\sim N\bigl(-\tfrac12\sigma_{\theta}^{2},\,\sigma_{\theta}^{2}\bigr)\)
  • Life‐cycle profile \(\Gamma\) from HFCS
  • Moments of stochastic process from NL micro data
    (de Nardi et al. 2021)
  • Two types: college vs. lower education
    • Different income levels and growth patterns
    • Different price/income ratios (but LtVs same)

Appendix

Data: saving rates in the HFCS

  • Saving rates increase with wealth for both

  • Decline in old age in NL
  • Interesting, as illiquidity of housing possible reason for plateau of saving (eg Yang 2009)

Back to main

Appendix

Data from Euro area countries

Saving rates over the wealth distribution (Q5 = 100)

Back to main

Appendix

Saving rates over the wealth distribution

Mortgaged homeowners vs. others

  • Waves 3 and 4:

Back to main

Appendix

Saving rates over the wealth distribution

Mortgaged homeowners vs. others

Back to main

Appendix

Flattening of saving rate differences reproduces pattern in the data

  • Saving rates increase for lower income (and younger ages)

Appendix

Saving rates over the life cycle

Saving by homeowners in NL and others

  • No substantial difference between post-policy reform mortgages

Back to main

Appendix

Saving rates over the wealth distribution

Saving by homeowners in NL and others

  • No substantial difference between post-policy reform mortgages

Back to main

Appendix

Age profiles of debt in the data

  • Life cycle profile of savings and mortgage debt
  • Strict subsample of households who:
    • Have never refinanced
    • Live in their first home
      • Roughly identified by age at purchase \(≤35\)
  • Interest-only mortgages: those for which amortization is \(<\) 80% implied by annuity formula

Appendix

Age profiles of debt in the data

Appendix

Age profiles of debt in the data

Appendix

Age profiles of debt in the data

Appendix

Model solution details

HH problem and solution

Basic principle uses stochastic gradient descent to find parameters of neural network that solve for the optimal policy function.

  • Machine learning techniques allow to compute the gradient \(\nabla_\theta \tilde{V}\left(s_0, \theta ; \hat{\pi}\right)\)
    • Computationally feasible with ML infrastructure, as neural networks are designed to work with problems with many dimension
    • JAX-based solution (implemented by Barrera & Silva, 2024, nndp)
    • Solved using Google Cloud TPU
  • Adjust \(\theta\) according to: \[ \Delta \theta = - \alpha \nabla_\theta \tilde{V}\left(s_0, \theta ; \hat{\pi}\right) \]
    • i.e., move in the direction that reduces the loss function (\(-V\)) the fastest
    • \(\alpha\) is the learning rate

Appendix

Bernstein and Koudijs (R&R QJE): mortgage amortization and saving in the Netherlands, 2012-16

Mortgage amortization & wealth accumulation in 2015 by first-time homeowners buying a house in 2012-13 Back to main